With the development of car industry, cars have been widely distributed to the extent that a household owns at least one car, and various high-tech electronic technologies are applied to the cars to improve safety of cars and promote driver's convenience.
Such high-tech electronic technologies include a vehicle surrounding image display system (e.g., Around View Monitoring (AVM)) for allowing a driver to easily check surroundings of a vehicle with his/her naked eyes by capturing surrounding of the vehicle and displaying a captured image. The vehicle surrounding image display system captures surroundings via cameras installed on the front, rear, left, and right of the vehicle, corrects an overlapping area to be seen naturally based on the captured images, and displays an image of the surroundings of the vehicle on a screen. Accordingly, the driver can exactly recognize the surroundings of the vehicle by viewing the image of the surroundings and can easily park the vehicle without viewing a side view mirror or a rear view mirror.
In particular, a 3-dimensional (3D) vehicle surrounding image providing system for providing an image of vehicle surroundings as a view image viewed from various viewpoints is in the spotlight. More specifically, techniques for enabling the driver to recognize the surroundings of the vehicle more easily by changing at least one of a location of a virtual viewpoint of a combined image, a direction of the eyes, and a focal distance according to a vehicle driving condition have been studied and developed.
In order to provide such a 3D image of vehicle surroundings, a general method maps the image captured by the camera to a 3D space virtual plane by using a 3D space model. In addition, the method creates a combined image based on the input image mapped to the 3D space virtual plane according to a corresponding relationship which is pre-defined according to a location of a camera virtual viewpoint, a direction of the eyes, and a focal distance which are defined according to a vehicle driving condition or a user selection. However, there is a problem that the combined image is not naturally seen according to the 3D space model. For example, when a hemisphere model may be used as the 3D space model, objects located at a long distance from the vehicle may be expressed naturally, but the ground surface around the vehicle is not expressed naturally. In addition, when the ground surface used in the related-art 2D vehicle surrounding image system is used as the mapping surface, objects other than the ground surface, for example, persons or obstacles standing around the vehicle have lengthened images and thus a normal image cannot obtained. In addition, since the view images are not well matched with real camera images during the matching process, there is a problem that it is difficult to match the images.